Rough set theory using similarity of objects described by ontology

Yusei Inukai, Andreas Gehrmann, Yoshimitsu Nagai, Syohei Ishizu

Abstract


Rough set theory was proposed by Pawlak Z. and is well used in the area of data mining. The main role of rough set theory is to extract important sets of attributes and decision rules based on the knowledge about objects. Rough set theory is defined by information system, whose role is knowledge representation. But recently the concept of ontology is used in knowledge engineering, Semantic Web, etc. Since ontology can flexibly represent knowledge, rough set theory using the concept of ontology enables us to use flexible information system in ontological description. One of our main aims of this paper is to propose rough set theory applied the concept of ontology to. For flexibility, ontology often consists of complex objects. We formulate a concept of similarity which measures a degree of relationship among complex objects. The concept of similarity is useful for extracting important sets of attributes defined by flexibly represented knowledge. We present steps for finding a set of decision rule based on the proposed concepts. We demonstrate the concepts and steps by using a simple example.

Keywords


Rough set theory, ontology, OWL, information system, degree of similarity

Full Text:

PDF